Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

voloctree: DDT Memory Failure when running with guard pages #302

Open
DMurataj01 opened this issue May 17, 2022 · 2 comments
Open

voloctree: DDT Memory Failure when running with guard pages #302

DMurataj01 opened this issue May 17, 2022 · 2 comments

Comments

@DMurataj01
Copy link
Contributor

Hi, I'm able to trigger the following memory error from using DDT, using the following settings using the following reproducer.

Looks like some sort of internal storage related issue but can you take a look and give your thoughts?

Thanks,
Denis

Screenshot 2022-05-17 at 18 38 32

Screenshot 2022-05-17 at 18 42 01

    void bitpit_mprotect_reproducer() {
        auto pabloBB = std::make_unique<bitpit::PabloUniform>(1000, 1000, 1000, 50, 3);

        // Set 2:1 balance for the octree.
        pabloBB->setBalanceCodimension(1);
        const int idx = 0;
        pabloBB->setBalance(idx, true);

        for(int i = 0; i < 6; ++i) {
            logInfo("Iteration: " + std::to_string(i));
            pabloBB->adaptGlobalRefine();
        }

        auto volTree = std::make_unique<bitpit::VolOctree>(std::move(pabloBB), &pabloBB);
        volTree->update(false);

        return {};
    }

Stack trace at point of failure:

Screenshot 2022-05-17 at 19 03 14

@andrea-iob
Copy link
Member

I tried you test case with both valgrind and gcc address sanitizer (active modules "undefined" and "address"), but I wasn't able to reproduce the problem.

The stack trace suggests that bitpit is failing when adding a new vertex (operator[] of an unordered_map).

Does the tests case run fine when run in release mode?

@DMurataj01
Copy link
Contributor Author

I tried you test case with both valgrind and gcc address sanitizer (active modules "undefined" and "address"), but I wasn't able to reproduce the problem.

The stack trace suggests that bitpit is failing when adding a new vertex (operator[] of an unordered_map).

Does the tests case run fine when run in release mode?

It runs but we consistently see this. Can you download a trial licence of ArmForge and run with guard pages on?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants